bussines_case_automation/clean_excel_template.py

#!/usr/bin/env python3
"""
Utility to clean Excel files from SharePoint/OneDrive metadata that causes
cross-platform compatibility issues.
"""
import os
import sys
import openpyxl
from pathlib import Path
import tempfile
import shutil


def clean_excel_file(input_path, output_path=None):
    """
    Clean an Excel file from SharePoint/OneDrive metadata.

    Args:
        input_path (str): Path to the input Excel file
        output_path (str): Path for the cleaned file (optional)

    Returns:
        bool: True if successful, False otherwise
    """
    if not os.path.exists(input_path):
        print(f"Error: File not found: {input_path}")
        return False

    if output_path is None:
        # Create cleaned version with _clean suffix
        path = Path(input_path)
        output_path = path.parent / f"{path.stem}_clean{path.suffix}"

    try:
        print(f"Loading Excel file: {input_path}")

        # Load workbook without VBA to avoid macro issues
        wb = openpyxl.load_workbook(input_path, data_only=False, keep_vba=False)

        # Clean metadata
        print("Cleaning metadata...")

        # Clear custom document properties
        if hasattr(wb, 'custom_doc_props') and wb.custom_doc_props:
            wb.custom_doc_props.props.clear()
            print("  ✓ Cleared custom document properties")

        # Clear custom XML
        if hasattr(wb, 'custom_xml'):
            wb.custom_xml = []
            print("  ✓ Cleared custom XML")

        # Clean core properties
        if wb.properties:
            # Keep only essential properties
            wb.properties.creator = "Excel Generator"
            wb.properties.lastModifiedBy = "Excel Generator"
            wb.properties.keywords = ""
            wb.properties.category = ""
            wb.properties.contentStatus = ""
            wb.properties.subject = ""
            wb.properties.description = ""
            print("  ✓ Cleaned core properties")

        # Create temporary file for double-save cleaning
        with tempfile.NamedTemporaryFile(suffix='.xlsx', delete=False) as tmp:
            tmp_path = tmp.name

        print("Saving cleaned file...")

        # First save to temp file
        wb.save(tmp_path)
        wb.close()

        # Re-open and save again to ensure clean structure
        print("Re-processing for maximum cleanliness...")
        wb_clean = openpyxl.load_workbook(tmp_path, data_only=False)

        # Additional cleaning on the re-opened file
        if hasattr(wb_clean, 'custom_doc_props') and wb_clean.custom_doc_props:
            wb_clean.custom_doc_props.props.clear()

        if hasattr(wb_clean, 'custom_xml'):
            wb_clean.custom_xml = []

        # Save final clean version
        wb_clean.save(output_path)
        wb_clean.close()

        # Clean up temporary file
        os.unlink(tmp_path)

        print(f"✓ Cleaned Excel file saved to: {output_path}")

        # Compare file sizes
        input_size = os.path.getsize(input_path)
        output_size = os.path.getsize(output_path)

        print(f"File size: {input_size:,} → {output_size:,} bytes")
        if input_size > output_size:
            print(f"Reduced by {input_size - output_size:,} bytes ({((input_size - output_size) / input_size * 100):.1f}%)")

        return True

    except Exception as e:
        print(f"Error cleaning Excel file: {e}")
        import traceback
        traceback.print_exc()
        return False


def clean_template():
    """
    Clean the template file in the template directory.
    """
    script_dir = os.path.dirname(os.path.abspath(__file__))
    template_dir = os.path.join(script_dir, 'template')

    # Look for template files
    possible_templates = [
        'Footprints AI for {store_name} - Retail Media Business Case Calculations.xlsx',
        'Footprints AI for store_name - Retail Media Business Case Calculations.xlsx'
    ]

    template_path = None
    for template_name in possible_templates:
        full_path = os.path.join(template_dir, template_name)
        if os.path.exists(full_path):
            template_path = full_path
            print(f"Found template: {template_name}")
            break

    if not template_path:
        print(f"Error: No template found in {template_dir}")
        return False

    # Create cleaned template
    cleaned_path = os.path.join(template_dir, "cleaned_template.xlsx")

    return clean_excel_file(template_path, cleaned_path)


if __name__ == "__main__":
    if len(sys.argv) > 1:
        # Clean specific file
        input_file = sys.argv[1]
        output_file = sys.argv[2] if len(sys.argv) > 2 else None

        if clean_excel_file(input_file, output_file):
            print("✓ File cleaned successfully")
        else:
            print("✗ Failed to clean file")
            sys.exit(1)
    else:
        # Clean template
        if clean_template():
            print("✓ Template cleaned successfully")
        else:
            print("✗ Failed to clean template")
            sys.exit(1)